Significant gaps in climate research in the Global South can lead to blind spots about the consequences of ongoing climate changes in these countries. New technologies, in particular the growing access to AI-based solutions able to run on personal computers, have the potential to improve the global monitoring of plankton communities. Part of my work is dedicated to developing simple solutions for plankton studies. I started out working on deep-learning based algorithms for plankton identification, and am growing increasingly interested in low-cost imaging devices which could easily be implemented in labs worldwide.
A. Deep-learning solutions for automatic planktonic identification
Convolutional neural networks are growing increasingly used in geosciences. Their use provides a means of generating standardised, reproducible datasets for plankton assemblages and size structure, and of increasing the spatial and temporal resolution of plankton studies.
Within the MANTA team at CEREGE, I am developing high throughput workflows for plankton detection and classifications, to be used for coastal biomonitoring and paleoceanographic studies.
In cases where classic thresholding techniques are ineffective, object detection workflows make it possible to extract and analyse individual plankton occurrences in microscope images. Figure from Godbillot et al., 2024.
B. Low-cost imaging techniques for coastal biomontoring
A 3D printer is used as a low-cost scanning device for the automatic acquisition of high resolution microfossil images.
Automated image acquisition systems for plankton studies remain sparse, in particular because they are prohibitively expensive and require taxonomic expertise for data analysis. New developments in robotics are now making it possible to develop affordable solutions for image acquisition, to be used on the field, or to be installed worldwide. Together with the growing number of community-annotated plankton image datasets, these new technologies have the potential to be installed globally for continuous monitoring. The simple solution for image acquisition presented here is being tested for its potential use in coastal monitoring of benthic foraminifera at the LPG in Angers.
References
2024
A New Method for the Detection of Siliceous Microfossils on Sediment Microscope Slides Using Convolutional Neural Networks
Camille Godbillot, Ross Marchant, Luc Beaufort, and 5 more authors
Journal of Geophysical Research: Biogeosciences, Sep 2024
Diatom communities preserved in sediment samples are valuable indicators for understanding the past and present dynamics of phytoplankton communities, and their response to environmental changes. These studies are traditionally achieved by counting methods using optical microscopy, a time‐consuming process that requires taxonomic expertise. With the advent of automated image acquisition workflows, large image data sets can now be acquired, but require efficient preprocessing methods. Detecting diatom frustules on microscope images is a challenge due to their low relief, diverse shapes, and tendency to aggregate, which prevent the use of traditional thresholding techniques. Deep learning algorithms have the potential to resolve these challenges, more particularly for the task of object detection. Here we explore the use of a Faster Region‐based Convolutional Neural Network model to detect siliceous biominerals, including diatoms, in microscope images of a sediment trap series from the Mediterranean Sea. Our workflow demonstrates promising results, achieving a precision score of 0.72 and a recall score of 0.74 when applied to a test set of Mediterranean diatom images. Our model performance decreases when used to detect fragments of these microfossils; it also decreases when particles are aggregated or when images are out of focus. Microfossil detection remains high when the model is used on a microscope image set of sediments from a different oceanic basin, demonstrating its potential for application in a wide range of contemporary and paleoenvironmental studies. This automated method provides a valuable tool for analyzing complex samples, particularly for rare species under‐represented in training data sets.
